Zero copy for GPUs

Second installment in AI/ML blogs

Understanding GPU Memory Organization Modern AMD GPUs, particularly the MI250X and MI300X based on the CDNA architecture, implement a deeply hierarchical memory system. Each level is optimized for...

🎤 Creating Slide Decks with Marp Marp (Markdown Presentation Ecosystem) allows you to write slide decks in plain Markdown and turn them into beautiful presentations. It’s simple, flexible, and int...

I spend most of my day working right in the terminal—whether I’m writing code, listening to music, or jotting down notes (you’re reading one of those notes right now). My setup is pretty classic: Z...

Introduction This entry is mostly from the perspective of learning the basic concepts involved in today’s buzz words Artificial intelligence and machine learning. Here, we cover some key milestone...

Generic design in C++ - policy based design

Smart Pointers in C++ Smart pointers in C++ help users manage dynamically allocated objects safely and efficiently. While raw pointers are inherently supported in C++, they come with several pitfa...

Recursion pattern

Here we will be discussing a recursive template for solving problems related to subsequence generation or power set generation. Generally speaking, this is a variant of the backtracking pattern, wh...